Recommendations for filter settings and sensitivity for various physiologic parameters
Channela
Low-frequency filter (Hz)
Time constant (s)
High-frequency filter (Hz)
Sensitivity
EEG
0.3
0.4
35
50 (µV/cm)
EOG
0.3
0.4
35
50 (µV/cm)
EMG
5b
0.03
90–120
20–50 (µV/cm)
ECG
1
0.12
15
1 MV/cm
Index of airflow
0.15
5b
15
c
Index of effort
0.15
5b
15
c
Bit Resolution and Sampling Rates
Unlike paper-based recording systems of the past that transformed the continuous analog signals into mechanical pen movements, contemporary PSG systems require analog-to-digital converters (ADCs) to convert the analog signals into digital form. The ADC converts the signals by assigning a numeric value to the amplitude of the analog waveforms at predetermined intervals. The number of binary units (bits) used to represent the numeric value of each sampled interval determines the amplitude resolution of the digital recording. For adequate amplitude resolution, a 12-bit or higher system is recommended.
The number of sampled intervals collected within the span of one second is defined as the sampling rate, which determines the frequency resolution of the recorded waveforms. The reconstruction of a digitized waveform follows the Nyquist theorem, which states that for basic frequency resolution the minimum sampling rate must be twice the rate of the highest frequency sampled (Nyquist rate). While the Nyquist rate is acceptable for processing signals that do not require detailed waveform reconstruction (such as the EMG), sampling rates that are significantly higher than the Nyquist rate are necessary for adequate graphic resolution of waveforms that require close scrutiny (such as EEG, EOG, and ECG waveforms) [25]. In these instances, the sampling rates should be at least 5–10 times higher than the highest frequency expected within the channel.
Unlike digital filtering, bit resolution and sampling rates must be applied during the recording process and cannot be changed afterward. To reduce overall file sizes and processing times, sampling rates can be selected channel by channel, depending on the type of signal recorded. Recommended sampling rates for various PSG parameters are included in Table 17.1.
Also to be considered is the display resolution, which is determined by the resolution of the monitor. The computer screen for review of the recording should have a sufficiently high resolution. Ideally, the screen should be at least 20 in. with a resolution of at least 1600 × 1200 pixels.
Time Scale
Historically, the speed of the chart drive for the recording instrument established the time scale and the epoch length (amount of time per page) of the recording. A common paper speed for traditional PSG was 10 mm/s, providing a 30-s epoch. Another widely accepted paper speed was 15 mm/s, a 20-s epoch length. For patients with suspected sleep-related seizure activity, a paper speed of 30 mm/s enhanced the ability to visualize EEG data. Data such as oxygen saturation and respiratory signals, however, were more easily visualized with slower paper speeds.
The issue of selecting the appropriate paper speed became moot when digital systems became the norm. In the present day, digital technology allows for multiple time-scale settings, which can be applied either during the recording or during playback (Figs. 17.1 and 17.2). For sleep stage and arousal scoring, the 30-s epoch continues to be the established standard [26]; however, alternative time scales can be used for evaluating sleep-related events. For example, an epoch length of 10–15 s can be useful for closely examining EEG or ECG abnormalities, whereas epoch lengths of 2–5 min are useful for evaluating respiratory patterns, periodic limb movements, and oximetry trends.
Fig. 17.1
PSG recording viewed in a 30-s epoch window. A time scale of 30 s per epoch provides a high level of detail in the EEG and EOG channels and is the standard scale for scoring sleep stages and arousals. This sample demonstrates the onset of stage 1 sleep. Reprinted with permission from Butkov [23]
Fig. 17.2
PSG recording viewed in a 5-min epoch window. After scoring sleep stages and arousals in 30-s epochs, the time scale can be compressed into 2- to 5-min epochs, to better discern respiratory patterns while correlating them to the patient’s sleep and wake physiology. This example shows a pattern of Cheyne–Stokes respiration emerging with the onset of NREM sleep. Reprinted with permission from Butkov [23]
The Study
Electrode and Sensor Application Process
The quality of the tracing generated in the sleep laboratory depends on the quality of the electrode application [27]. Before any electrode or sensor is applied, the patient should be instructed about the procedure and given an opportunity to ask questions. The first step in the electrode application process involves measurement of the patient’s head. The International 10–20 system [28] of electrode placement is used to localize specific electrode sites (Figs. 17.3 and 17.4) (see also Chap. 24). The following sections address the application process for EEG, EOG, EMG, and ECG electrodes.
Fig. 17.3
International 10/20 system of electrode placement (top view). Reprinted with permission from Butkov [23]
Fig. 17.4
International 10/20 system of electrode placement (side view). Reprinted with permission from Butkov [23]
Electroencephalography
As originally described in the Rechtschaffen and Kales manual [29], standard electrode derivations for monitoring EEG activity during sleep are C3/A2 or C4/A1 for central EEG activity, and O1/A2 or O2/A1 for occipital EEG activity (see also Chap. 18). The AASM Scoring Manual terminology uses the terms “M1” and “M2” instead of “A1” and “A2” for the reference electrodes placed on the mastoid process. In this case, the derivations are C3/M2 or C4/M1, and O1/M2 or O2/M1 (the AASM recommends using C4/M1 and O2/M1 as the default derivations, with C3/M2 and O1/M2 as alternative back-up derivations). The AASM also recommends the use of frontal EEG recordings (F4/M1 and F3/M2) when considering decisions regarding K complexes or slow-wave activity [26].
In some situations, there may be a need for additional electrodes. For example, to rule out the possibility of epileptic seizures during sleep or to detect the presence of other sleep-related EEG abnormalities, it may be necessary to apply the full complement of EEG electrodes according to the International 10–20 system. An abbreviated montage to screen for EEG abnormalities during PSG is discussed in Appendix 17.3. For recording EEG, gold cup electrodes are commonly used. The electrode cups are filled with conductive cream or electrode paste and adhered to the scalp with collodion or with additional electrode paste. The collodion technique [27] has long been an accepted and preferred method of application for EEG scalp and reference electrodes. Other methods using electrode paste and a conductive medium are acceptable and may be preferred in certain conditions.
The International 10–20 system of electrode placement determines the placement of EEG electrodes. Reference electrodes are placed on the bony surface of the mastoid process. A description of the measurement procedure appears in Appendix 17.4.
Electrooculography
The EOG is a recording of the movement of the corneoretinal potential difference that exists in the eye. It is the movement of this dipole with respect to the placement of the EOG electrodes that is recorded (Fig. 17.5) (see also Chap. 18). Gold cup electrodes or silver–silver chloride electrodes can be used to record the EOG. EOG electrodes are typically applied to the surface of the skin with an adhesive collar; this method avoids the risk of collodion contacting the patient’s eyes.
Fig. 17.5
Recording montage for a 2-channel EOG demonstrates out-of-phase deflections in the EOG tracings associated with conjugate eye movements. The deflections in the EOG tracings follow the standard EEG and PSG polarity convention (negative voltage = upward deflection; positive voltage = downward deflection)
The AASM Scoring Manual recommends placing the right EOG electrode (E2) one centimeter directly above the right outer canthus (ROC) and the left EOG electrode (E1) one centimeter directly below the left outer canthus (LOC) [26]. Some adjustments might be necessary to these placements, to avoid placing the E2 electrode directly over the right eyebrow or to avoid placing either of the electrodes too close to the sensitive tissue of the eyelid. The AASM also recommends referencing both EOG electrodes to the right mastoid (M2); however, a contralateral reference (E2/M1 and E1/M2) may be preferable for maximizing signal amplitudes in both EOG channels and equalizing the out-of-phase signals deflections seen with conjugate eye movements [30].
It should be noted that many variations of electrode placement and recording derivations have been used in a variety of clinical and research settings. Additional infraorbital and supraorbital electrodes enhance the ability to detect eye movements that occur in the vertical plane and can be particularly useful in the MSLT [31, 32]. Given the existing variations in methodology within the clinical and research environments, it is important to know the exact electrode placements and inputs to the EOG channels when interpretation of EOG activity has significant impact on diagnosis or treatment outcome (Fig. 17.6).
Fig. 17.6
Schematic diagram showing the placement of the EOG, chin EMG, patient ground and system reference electrodes
Electromyography
To record chin EMG activity, gold cup or silver–silver chloride electrodes are placed over the mentalis and submentalis muscles. The electrodes can be attached with double-sided electrode collars and paper tape, or they can be adhered with collodion or paste if the patient has a beard. The AASM Scoring Manual recommends placing one electrode on the midline, 1 cm above the inferior edge of the mandible, and two electrodes 2 cm below the inferior edge of the mandible, offset 2 cm to the right and left of midline, respectively [26]. Two of the electrodes are used to create a bipolar EMG recording, while the third electrode serves as a backup. Some practitioners place an electrode over a masseter muscle to better detect bursts of EMG activity associated with bruxism (see Fig. 17.6).
To record leg EMG, a pair of electrodes is placed over the anterior tibialis muscle of each leg. The electrodes are spaced approximately 2–3 cm apart and referenced to each other in a bipolar derivation.
Electrocardiography
There are a variety of approaches for recording the ECG during PSG. The simplest approach is to use disposable ECG patches with a stress loop incorporated into the lead wire to ensure long-term placement. A modified lead II is obtained by placing an electrode slightly below the right clavicle and referencing it to an electrode placed over the left lower thorax. A third electrode placed slightly below the left clavicle allows for recording a modified lead I (left subclavicular to right subclavicular) and lead III (left subclavicular to left lower thorax) ECG.
Patient Ground and System Reference Electrodes
A single patient ground electrode is applied typically to the patient’s forehead, slightly below the hair line. The purpose of a patient ground is to divert excessive 50 or 60 Hz line frequency interference from the patient’s body. The input for the ground electrode must be isolated to prevent the possibility of stray current passing through the patient, and only a single ground electrode should be used on the patient to avoid the possibility of creating a ground loop.
If system referencing capabilities are provided by the PSG equipment, then an additional system reference electrode is attached to the patient, typically placed on the midline of the scalp (Cz). System referencing offers a way to select or to change input signal derivations either during data collection or during playback. During data collection, the signals from all of the applied electrodes are initially referenced to the Cz electrode. This configuration is not seen by the operator, but is used as a framework for selecting the desired derivations. For each selected derivation, the computer subtracts the common reference (Cz) from the chosen pair of input signals and displays the resulting derivation on the computer monitor [23].
It is important to note that a faulty connection to either the patient ground or the system reference electrode will result in signal degradation that can affect all of the PSG channels. A faulty patient ground connection can cause excessive 50 or 60 Hz interference in the recording, while a faulty system reference connection can cause complete signal loss in all channels (with the exception of any dedicated bipolar or transduced signals that bypass the common reference). Consequently, extra care should be taken to ensure that these two electrodes are properly applied and well maintained throughout the duration of the recording.
Electrode Impedances
Before recording, all electrodes should be visually inspected to check the security of their placement and an impedance check should be performed and documented. An impedance meter is ideally part of the recording system. Alternatively, a separate device can be used. Adjustments should be made to any electrode with impedance readings of greater than 5000 Ω (5 kΩ). Impedance levels are reduced by carefully cleansing and scrubbing each electrode site with a gel-based skin prepping solution before applying the electrode. Only a small area of skin should be scrubbed, no larger than the size of the electrode cup, to prevent electrical bridging and to minimize the occurrence of artifacts in the recording [23].
Physiologic Calibrations
Physiologic calibrations are performed after the electrode and sensor application is complete. This calibration documents proper functioning of the electrodes and other recording devices and provides baseline data for review and comparison when scoring the PSG. The specific instructions given to the patient for this calibration include
Eyes open, look straight ahead for 30 s.
Eyes closed, look straight ahead for 30 s.
Hold head still, look to left and right, up and down. Repeat.
Hold head still, blink eyes slowly, five times.
Clench the jaw and grit the teeth.
Inhale and exhale slowly, three times.
Hold breath for 10 s.
Dorsiflex right foot (i.e., bend the foot upward), dorsiflex left foot (this ensures contraction of the tibialis anterior muscle).
As these instructions are given to the patient, the technologist examines the tracings and documents the patient’s responses. When the patient stares straight ahead for 30 s with eyes open, the background EEG activity is examined. As the patient looks right and left, the tracing is examined for out-of-phase deflections of the signals associated with recording the EOG. Out-of-phase deflection occurs if the inputs to consecutive channels of the polygraph are E2/M1 for the first EOG channel and E1/M2 for the second. It is also important, when the patient closes his or her eyes, to observe the reactivity of the alpha rhythm seen most prominently in the occipital EEG; alpha rhythm is usually best visualized when the patient’s eyes are closed.
The mentalis/submental EMG signal is checked by asking the patient to clench the jaws, grit the teeth, or yawn. The technologist documents proper functioning of the electrodes and amplifiers used to monitor anterior tibialis EMG activity by asking the patient to dorsiflex the right foot and the left foot in turn. If rapid eye movement (REM) sleep behavior disorder is suspected, additional electrodes should be applied to the surface of the skin above the extensor digitorum communis muscles of each arm. Patients are asked to extend their wrists while the technologist examines the recording for the associated increase in amplitude in the corresponding EMG channel.
Inspiration and expiration allow for examination of the channels monitoring airflow and respiratory effort [33, 34]. A suggested convention is that inspiration causes an upward deflection of the signal and expiration a downward deflection. It is important that the signals in all the channels recording respiration are in phase with each other to avoid confusion with paradoxical breathing. The technologist should observe a flattening of the trace for the duration of a voluntary apnea. [Note: It is recommended to include end-tidal or transcutaneous CO2 monitoring when studying children [35], or adult patients with underlying lung or neuromuscular diseases. The addition of CO2 monitoring increases the sensitivity of the study of hypoventilation.
Fig. 17.7
60 Hz artifact. This is an example of 60 Hz artifact in the chin EMG channel caused by faulty electrode connection. Reprinted with permission from Butkov [23]
If the 50 or 60 Hz notch filter (also called “line filter” or “AC filter”) is in use, a brief examination (2–4 s) of portions of the tracing with the filter in the “out” position is essential. This allows for identification of any 50 or 60 Hz interference that may be masked by the filter. Care should be taken to eliminate any source of interference and to ensure that the 50 or 60 Hz notch filter is used only as a last resort. This is most important when recording patients suspected of having seizure activity, because the notch filter attenuates the amplitude of the cortical spike activity seen in association with epileptogenic activity.
The physiologic calibrations enable the technologist to determine the quality of data before the PSG begins. If artifacts are noted during the physiologic calibrations, it is imperative that every effort be made to correct the problem, as the condition is likely to get worse through the remaining portions of the recording. The functioning of alternative backup electrodes should also be examined during this calibration. If any additional monitoring devices are used for the study, the technologist should incorporate the necessary calibrations.
When a satisfactory calibration procedure and all other aspects of patient and equipment preparation are completed, the patient is told to assume a comfortable sleeping position and to try to fall asleep. Then, the lights are turned out in the patient’s room and the “lights-out” time is noted on the tracing and in the recording log.
Monitoring and Recording
Complete documentation for the PSG is essential. This includes patient identification (patient’s full name and medical record number), date of recording, and a full description of the study. The name of the technologist performing the recording, as well as those of any technologists who prepared the patient or the equipment, should be noted. In laboratories that use multiple pieces of equipment, the specific instrument used to generate the recording should be identified. This is particularly useful in the event that artifacts are noted during the analysis portion (scoring) of the sleep study.
Specific parameters recorded on each channel should be clearly identified, as should a full description of sensitivity, filter, and calibration settings for each channel. The time of the beginning and end of the recording must be recorded, as well as specific events that occurred during the night. Any changes made to filter, sensitivity, or derivation settings during the recording should be clearly noted in the study.
The technologist is also responsible for providing a clinical description of unusual events. For example, if a patient experiences an epileptic seizure during the study, the clinical manifestations of the seizure must be detailed: deviation of eyes or head to one side or the other, movement of extremities, presence of vomiting or incontinence, duration of the seizure, and postictal status. Similar information should be reported on any clinical event observed in the laboratory, such as somnambulism or clinical features of REM sleep behavior disorder. Physical complaints reported by the patient are also noteworthy. When studying patients with SRBD, documentation of snoring, wheezing, labored breathing, or other descriptive observations of the patient’s breathing patterns provide essential information that can assist the reading physician in interpreting the PSG recording.
Troubleshooting and Artifact Recognition
In general, when difficulties arise during recording, the troubleshooting inquiry begins with the patient and follows the path of the signal to the recording device. More often than not, the problem can be identified as a displaced or faulty electrode or sensor. It is less likely that artifact is the result of a problem with an amplifier. If the artifact is generalized (i.e., in most channels), then the integrity of the ground electrode, system reference electrode, or the cable from the electrode jack box should be checked. If the artifact is localized (i.e., in a limited number of channels), then the question should be: which channels have this artifact in common and what is common to the channels involved? The artifact is probably the result of a problem with an electrode or sensor that is common to both channels. If the artifact is isolated to a single channel, then the problem most likely stems from the specific electrode or sensor used only for that channel. In many instances, localized artifacts can be corrected by changing the derivation(s) to an alternative reference electrode, or by using alternative backup derivations.
50 or 60 Hz Artifact
50 or 60 Hz interference (Figs. 17.7 and 17.8) stems from power line frequency in the vicinity of the study.1 As described earlier in this chapter, common mode rejection (a function of the PSG differential amplifiers) serves to eliminate 50/60 Hz artifact from the PSG recording. In addition, the patient ground electrode is used to divert stray electrical interference from the patient. The effectiveness of common mode rejection and the function of the patient ground are both dependent on the integrity of the electrode connections to the patient. Consequently, the presence of 50 or 60 Hz artifact is most often caused either by high electrode impedances (which defeat the function of common mode rejection), or by a faulty ground connection (which prevents effective diversion of electrical interference from the patient), or a combination of both. Such 50 or 60 Hz interference can also be aggravated by excessive leakage current, which can stem from any electrical equipment in the vicinity of the study.
